CN107333031A - A kind of automatic edit methods of multi-channel video suitable for campus football match - Google Patents
A kind of automatic edit methods of multi-channel video suitable for campus football match Download PDFInfo
- Publication number
- CN107333031A CN107333031A CN201710623659.6A CN201710623659A CN107333031A CN 107333031 A CN107333031 A CN 107333031A CN 201710623659 A CN201710623659 A CN 201710623659A CN 107333031 A CN107333031 A CN 107333031A
- Authority
- CN
- China
- Prior art keywords
- background
- video
- pixel
- image
- moving target
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N5/00—Details of television systems
- H04N5/222—Studio circuitry; Studio devices; Studio equipment
- H04N5/2224—Studio circuitry; Studio devices; Studio equipment related to virtual studio applications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N5/00—Details of television systems
- H04N5/222—Studio circuitry; Studio devices; Studio equipment
- H04N5/262—Studio circuits, e.g. for mixing, switching-over, change of character of image, other special effects ; Cameras specially adapted for the electronic generation of special effects
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Image Analysis (AREA)
- Television Signal Processing For Recording (AREA)
- Closed-Circuit Television Systems (AREA)
Abstract
The present invention relates to a kind of automatic edit methods of multi-channel video suitable for campus football match, including：Video acquisition：Shot by multichannel video camera and precincts of the bath demarcation, obtain accurate multiple paths of video images；Video is edited automatically：Video background modeling and ball, people's moving object detection are passed sequentially through to described multiple paths of video images, multi-channel video is compiled as automatically to export video all the way.Compared with prior art, the present invention have it is cost-effective, the advantages of simple to operate, convenient.
Description
Technical field
The present invention relates to a kind of section of football match video edit methods, more particularly, to a kind of suitable for campus football match
The automatic edit methods of multi-channel video.
Background technology
In the football activity of campus, the video editing of football match is the important technology for supporting campus sports and teaching
One of.For the football match of professional class, its video content usually relies on multigroup, set of professional imaging device and is acquired,
Completed followed by a large amount of human-editeds, this video acquisition is with edit mode mostly just for an important match.With
On the contrary, the video data volume of campus football it is big, mostly collection from inexpensive amateur picture pick-up device, be difficult to adopt substantial amounts of people
Work edit mode.
With the development of image processing techniques, the automatic analysis technology of video is more and more applied to section of football match video
Editor.Such as augmented reality, sportsman track automatic analysis technology etc..But these technologies are to data acquisition equipment, court
Environment etc. has compared with strict requirements, is only applicable to the professional race venue of high standard, and the applicability for campus football is relatively low.
Such as many technologies are required at the top of court setting up high resolution camera, and this is difficult to realize for common campus place
's.
How in inexpensive hardware device, realize the editor of football video for campus under the conditions of seldom manual intervention
Had very important significance for football activity.
The content of the invention
It is an object of the present invention to overcome the above-mentioned drawbacks of the prior art and provide one kind is applied to campus foot
The automatic edit methods of multi-channel video of ball match.
The purpose of the present invention can be achieved through the following technical solutions：
A kind of automatic edit methods of multi-channel video suitable for campus football match, including：
Video acquisition：Shot by multichannel video camera and precincts of the bath demarcation, obtain accurate multiple paths of video images；
Video is edited automatically：Video background modeling and ball, people's motion mesh are passed sequentially through to described multiple paths of video images
Mark detection, multi-channel video is compiled as automatically to export video all the way.
Described video acquisition is concretely comprised the following steps,
1) video camera frame per second is adjusted, video camera is opened and is set to video shooting mode；
2) stopwatch of same Millisecond is shot with four video cameras；
3) keep video camera to be shooting state, four video cameras are set up into four angle points in court respectively；
4) keep the camera angle of every video camera constant, carry out video capture；
5) No. four video cameras are carried out by time synchronized according to the stopwatch two field picture initially shot, finds out each moment corresponding
Four road images；
6) the court lawn scope in four road image frames is marked, separates court lawn space and non-lawn region；
7) lawn region in image is carried out after background estimating, output multi-channel video image.
What described video was edited automatically concretely comprises the following steps,
1) background modeling and analysis are carried out to court lawn space image, obtains the moving target UNICOM body in four road images；
2) all moving target UNICOMs body is detected, football is calibrated；
3) automatic decision captures the video camera of football, contains football if only having in image all the way, Ze Jianggai roads image is set
For the present frame in video clipping；If containing football in multiway images, compare area of the football in each image, foot will be contained
The maximum image of ball image area is set to present frame；If containing the difference between football, and image area in multiway images 10%
Within, then judged according to the gross area of all moving target UNICOMs body in image, by the moving target UNICOM body gross area most
Big image is set to the present frame of video frequency output；
4) optimum image among four tunnels of output is as current video frame, the video after being edited.
Described background modeling and analysis include carrying out 100 frame video images initially recorded successively initialization background and
Context update.
Described initialization background is concretely comprised the following steps,
1) obtain frame by frame per two field picture, and record the pixel per frame, if the image of current record is 100 frames, carry out next
Step；If less than 100 frames, continuing to read until equal to 100 frames, carrying out next step；
2) pixel of 100 frame video images to initially recording is classified as three classes in the way of C mean clusters, each
Class is considered as a class background, count the average and standard deviation of all pixels color of each classification background as the central value of background and
Excursion, and record the number of pixels for belonging to every class background；
3) differential threshold of a class center value is set, if the Euclidean distance of certain two class background central value is less than the threshold
Value, then merge two class backgrounds.
Described context update concrete mode is,
1) according to the pixel and its color value of newest acquisition, calculate respectively between the color value and background central point of all categories
Euclidean distance, the background where minimum value is considered as and the immediate background of new pixel；
If 2) be less than 10, and new pixel color value and the closest back of the body with the current pixel quantity of the immediate background of new pixel
The Euclidean distance of the average of scape is less than 20, then is directly judged as that new pixel belongs to background, new pixel is added into this classification background
In, and make the quantity of pixel plus 1；If the pixel quantity after updating is more than or equal to 10, the central value and standard of the background are counted
Difference, deletes record time earliest pixel in the total pixel of background, it is ensured that the total pixel number amount of three class backgrounds is n；
If 3) be more than or equal to 10 with the current pixel quantity of the immediate background of new pixel, and new pixel color value with most
Euclidean distance close to the average of background is less than poor 3 times of background standard, then new pixel is added in this classification background, and more
The central point and standard deviation of this new classification background, delete the earliest pixel of record time in the total pixel of background, it is ensured that background
Total pixel number amount is n, and wherein n is 100；
If 4) be not belonging to two kinds of situations above with the immediate background of new pixel, then it is assumed that be found that the back of the body of new category
Scape；If the background quantity at current time is equal to 3, that minimum class background of pixel quantity is deleted, and deletion belongs to the back of the body
The all pixels of scape；After deletion, new classification background is set up by new pixel, new background central value is the color value of new pixel,
The pixel quantity of new background is 1, and standard deviation wouldn't be estimated.
The acquisition process of described moving target UNICOM body is that the new pixel for each frame occur is used as moving target candidate
Pixel；For the moving target candidate pixel in each frame, the combination of pixels that all eight neighborhoods are connected is constituted many with this
Individual moving target UNICOM body.
The operation of described all moving target connected components of detection is, according to the area of described moving target UNICOM body and
The current location of football is judged with round degree of approximation, and forward sight is worked as according to the areal calculation of described moving target UNICOM body
The size of moving target in frequency.
The frame per second adjusting range of described video camera is 30-60.
The antenna height scope of described video camera is 2~4m.
Compared with prior art, the present invention has advantages below：
1st, it is cost-effective：Video is completed using the inexpensive video camera in 4 tunnels to edit automatically, without setting up energy directly over court
Enough cover the high-resolution professional camera of the whole audience.
2nd, it is simple to operate, convenient：Utilize the demarcation of video camera precincts of the bath, video background modeling and ball, the inspection of people's moving target
4 tunnel video editings are automatically output result video all the way by survey mode, and automaticity is high, it is not necessary to manual intervention.
Brief description of the drawings
Fig. 1 for the present invention in the automatic edit methods of video flow chart；
Fig. 2 is the background modeling and algorithm of target detection flow chart in video editing part of the present invention；
Fig. 3 is camera position layout diagram of the invention.
Embodiment
Below in conjunction with the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is carried out clear, complete
Site preparation is described, it is clear that described embodiment is a part of embodiment of the present invention, rather than whole embodiments.Based on this hair
Embodiment in bright, the every other reality that those of ordinary skill in the art are obtained on the premise of creative work is not made
Example is applied, should all belong to the scope of protection of the invention.
The present invention is applied to the automatic edit methods of multi-channel video of campus football match.This method is divided into two parts：1) regard
Frequency is obtained；2) video is edited automatically.
In video acquisition step, prepare four inexpensive video cameras (such as mobile phone, common camera) first, by video
Frame rate is adjusted to 30-60 frames.First cameras is opened and enters video shooting mode, and is shot with four video cameras
The stopwatch of same Millisecond.Then camera function is not closed, camera pedestal is set to four angle points in court.Shoot stopwatch
Frame of video is in follow-up analysis by the time synchronized for video.In shooting process, the shooting angle of every video camera is all
It is fixed.The antenna height of video camera may be selected 2 to 4 meters.
In the automatic editor of video, by automatic decision that all the way video camera can capture football, the road images of Bing Jiang tetra-
In the present frame for being set to output result video all the way.
First according to the shooting angle of every video camera, the Football Field Turf range flags in picture are come out, i.e., to drawing
Face sets a quadrangle, and quadrangle interior zone is court lawn, and quadrangle perimeter is non-lawn region.
During video acquisition, background estimating is carried out to lawn region in image first.
For N (N=100) frame video image of most start recording, background is initialized as follows:
A) for the pixel of N two field pictures, 3 classes, all pixels system of each class are classified as in the way of C mean clusters
The mean μ and standard deviation sigma of its color are counted as central value and excursion, and records the number of pixels Si for belonging to every class background.
B) the differential threshold T of one class center value is set again, if two classes central value Euclidean distance be less than T if
Two categories combinations.Therefore at each moment, the background classification of each pixel is up to 3 classes, minimum 1 class.
For the pixel p of newest acquisition, its color value is Color, calculates central point and the immediate background classes of its color
Other i.And background is updated according to following three kinds of modes.
If a) be less than 10 with the immediate background i of new pixel current pixel quantity Si, and new pixel color with most
Euclidean distance close to the average of background is less than 20, then directly is judged as new pixel to belong to background i.And be added to new pixel
Among background i, and Si is led plus 1.If the Si after updating is more than or equal to 10, the central value and standard deviation of the background are counted.This
That pixel of record time earliest in the total pixel of background is deleted afterwards, it is ensured that the total pixel number amount of three class backgrounds is N.
If b) the current pixel quantity Si with the immediate background i of new pixel is more than or equal to 10, new pixel is calculated
Color Color and background classification i center μiEuclidean distance d (p-i)=| Color- μi|, if d (p-i) is less than background i's
3 times of standard deviation sigmai, then new pixel is added in the category, and update the central point μ of the categoryiAnd standard deviation sigmai.Hereafter delete
Except that pixel that the record time in the total pixel of background is earliest, it is ensured that the total pixel number amount of background is N.
If c) being not belonging to two kinds of situations above with the immediate background i of new pixel, then it is assumed that be found that new background.
Now, if current background quantity is equal to 3, that minimum background of pixel quantity Si is deleted, and deletion belongs to the background
All pixels.After deletion, new background classification is set up with new pixel, background classification central value is the color Color of new pixel, newly
The pixel quantity of background is 1, and standard deviation wouldn't be estimated.
During context update, the new pixel occurred in each frame is recorded, the candidate pixel of moving target is used as.For
Moving target candidate pixel in each frame, the combination of pixels that all 8 neighborhoods are connected is got up, and constitutes multiple UNICOM's bodies.According to
The area of UNICOM's body and the current location that football is judged with round degree of approximation.Further according to the areal calculation current video of UNICOM's body
The size of middle moving target.
Time synchronized is carried out according to the stopwatch picture frame for most starting to shoot to thinking video camera, each moment is found corresponding
Four road images.
Background modeling and analysis are all carried out to four road images, the moving target UNICOM body in four tunnels is obtained.And according to as follows
Mode sets the present frame of video editing result：
Circularity highest UNICOM body in UNICOM's body is found, is demarcated as football.
If only having football in image all the way, then the road image is just set to the present frame in video clipping.
It is if having football in multiway images, then compare the area of football in the picture, football image area is maximum
Image be set to present frame.
If have football in multiway images, and its image area difference within 10%, then according in image own
The gross area of moving target judges.The bigger image of the moving target gross area is set to video frequency output present frame.
Fig. 1 show the overview flow chart of the automatic edit methods of multi-channel video of the present invention, and key step has：
Step 1：Video time is completed to the video image that No. 4 video cameras are shot synchronous.Prepare four low costs first to take the photograph
Camera (such as mobile phone, common camera), 30-60 is adjusted to by video capture frame per second.Open video camera and be set to video bat
Pattern is taken the photograph, the stopwatch of same Millisecond is shot using this four video cameras.Holding video camera is shooting state, by camera pedestal
Four angle points in court are set to, as shown in Figure 3.The frame of video for shooting stopwatch will be same for the time of video in follow-up analysis
Step.In shooting process, the shooting angle of every video camera is all fixed.The antenna height of video camera may be selected 2 to 4
Rice.
Step 2：Pair according to the shooting angle of every video camera, the Football Field Turf range flags in picture are come out, i.e.,
Picture sets a quadrangle, and quadrangle interior zone is court lawn, and quadrangle perimeter is non-lawn region.
Step 3：To carrying out background modeling and analysis per video all the way.
Step 4：Moving object detection, i.e., to the new pixel occurred in each frame, i.e., the non-back of the body are carried out using the background of modeling
Scene element, is used as the candidate pixel of moving target.The moving target candidate pixel that 8 neighborhoods are connected is combined, constituted multiple
UNICOM's body.The current location of football is judged according to the area of UNICOM's body and with round degree of approximation, further according to the area of UNICOM's body
Calculate the size of moving target in current video.
Step 5：Using moving target (including the ball and sportsman) area of UNICOM detected, circularity highest in area of UNICOM is found
UNICOM's body, is demarcated as football.If only having football in image all the way, then just the road image is set in video clipping
Present frame.If having has football in multiway images, then compare the area of football in the picture, by the figure that football image area is maximum
As being set to present frame.If have football in multiway images, and its image area difference within 10%, then according to institute in image
There is the gross area of moving target to judge.The bigger image of the moving target gross area is set to video frequency output present frame.
In the step 3 of above-mentioned flow, the background modeling flow of 4 road videos is as shown in Figure 3 in detail：
Step 1：Obtain frame by frame per two field picture, pixel of the record per frame.If the amount of images of current record is less than 100 frames,
Then continue to read, until equal to 100 frames, carrying out step 2.1.
Step 2.1：C mean clusters are carried out using 100 pixels of record, 3 classes are classified as.
Step 2.2：The color mean μ and standard deviation sigma of each class background are recorded as central value and excursion, and is recorded
Belong to the number of pixels Si of every class background.The differential threshold T=10 of one class center value is set again, if the central value of two classes
Euclidean distance is less than T, then by two categories combinations.Therefore at each moment, the background classification of each pixel is up to 3 classes, most
Low is 1 class.
Step 3：For the pixel p of newest acquisition, its color value is Color, calculates central point and its color is immediate
Background classification i (optimal background).If the pixel quantity of optimal background is less than 10, and optimal background mean value and the color of newest pixel
Euclidean distance d (p-i)=| Color- μ i | less than 20, jump to step 4.If the pixel quantity of optimal background is more than 10, and
Optimal background mean value and the Euclidean distance d (p-i) of the color of newest pixel are less than 3 times of optimal background standard difference, jump to step
Rapid 4.Other situations then jump to step 5.1.
Step 4：New pixel is added among background i, and optimal background i current pixel quantity Si is added 1.If updating
Si afterwards is more than or equal to 10, recalculates the central value and standard deviation of the background.Delete in the total pixel of background and record after calculating
That pixel of time earliest, it is ensured that the total pixel number amount of all categories background is 100.Background modeling flow terminates.
Step 5.1：If current background quantity is 3, the minimum background of pixel quantity is deleted, and deletion belongs to the background
All pixels.
Step 5.2：New background classification is set up with new pixel, background classification central value is the color Color of new pixel, newly
The pixel quantity of background is 1, and standard deviation wouldn't be estimated.
Step 5.3：By the candidate pixel that emerging pixel record is moving target.
Step 5.4：For the moving target candidate pixel in each frame, the combination of pixels that all 8 neighborhoods are connected is got up,
Constitute multiple UNICOM's bodies.The current location of football is judged according to the area of UNICOM's body and with round degree of approximation.Further according to UNICOM
The size of moving target in the areal calculation current video of body.
The foregoing is only a specific embodiment of the invention, but protection scope of the present invention is not limited thereto, any
Those familiar with the art the invention discloses technical scope in, various equivalent modifications can be readily occurred in or replaced
Change, these modifications or substitutions should be all included within the scope of the present invention.Therefore, protection scope of the present invention should be with right
It is required that protection domain be defined.
Claims (10)
1. a kind of automatic edit methods of multi-channel video suitable for campus football match, it is characterised in that including：
Video acquisition：Shot by multichannel video camera and precincts of the bath demarcation, obtain accurate multiple paths of video images；
Video is edited automatically：Video background modeling and ball, the inspection of people's moving target are passed sequentially through to described multiple paths of video images
Survey, multi-channel video is compiled as automatically to export video all the way.
2. a kind of automatic edit methods of multi-channel video suitable for campus football match according to claim 1, its feature
It is, described video acquisition is concretely comprised the following steps,
1) video camera frame per second is adjusted, video camera is opened and is set to video shooting mode；
2) stopwatch of same Millisecond is shot with four video cameras；
3) keep video camera to be shooting state, four video cameras are set up into four angle points in court respectively；
4) keep the camera angle of every video camera constant, carry out video capture；
5) No. four video cameras are carried out by time synchronized according to the stopwatch two field picture initially shot, finds out corresponding four tunnel of each moment
Image；
6) the court lawn scope in four road image frames is marked, separates court lawn space and non-lawn region；
7) lawn region in image is carried out after background estimating, output multi-channel video image.
3. a kind of automatic edit methods of multi-channel video suitable for campus football match according to claim 1, its feature
It is, what described video was edited automatically concretely comprises the following steps,
1) background modeling and analysis are carried out to court lawn space image, obtains the moving target UNICOM body in four road images；
2) all moving target UNICOMs body is detected, football is calibrated；
3) automatic decision captures the video camera of football, contains football if only having in image all the way, Ze Jianggai roads image is set to regard
Present frame in frequency editing；If containing football in multiway images, compare area of the football in each image, football figure will be contained
The maximum image of image planes product is set to present frame；If containing the difference between football, and image area in multiway images within 10%,
Then judged according to the gross area of all moving target UNICOMs body in image, by the figure that the moving target UNICOM body gross area is maximum
Present frame as being set to video frequency output；
4) optimum image among four tunnels of output is as current video frame, the video after being edited.
4. a kind of automatic edit methods of multi-channel video suitable for campus football match according to claim 3, its feature
It is, described background modeling and analysis includes carrying out initialization background and the back of the body successively to 100 frame video images initially recorded
Scape updates.
5. a kind of automatic edit methods of multi-channel video suitable for campus football match according to claim 4, its feature
It is, described initialization background is concretely comprised the following steps,
1) obtain frame by frame per two field picture, and record the pixel per frame, if the image of current record is 100 frames, carry out next step；If
Less than 100 frames, then continue to read until equal to 100 frames, carrying out next step；
2) pixel of 100 frame video images to initially recording is classified as three classes in the way of C mean clusters, and each class is regarded
For a class background, the central value and change as background of average and standard deviation of all pixels color of each classification background is counted
Scope, and record the number of pixels for belonging to every class background；
3) differential threshold of a class center value is set, if the Euclidean distance of certain two class background central value is less than the threshold value,
Two class backgrounds are merged.
6. a kind of automatic edit methods of multi-channel video suitable for campus football match according to claim 4 or 5, it is special
Levy and be, described context update concrete mode is,
1) according to the pixel and its color value of newest acquisition, the Europe between the color value and background central point of all categories is calculated respectively
Family name's distance, the background where minimum value is considered as and the immediate background of new pixel；
If 2) be less than 10 with the current pixel quantity of the immediate background of new pixel, and new pixel color value with closest to background
The Euclidean distance of average is less than 20, then is directly judged as that new pixel belongs to background, new pixel is added in this classification background, and
The quantity of pixel is made plus 1；If the pixel quantity after updating is more than or equal to 10, the central value and standard deviation of the background are counted, is deleted
Except the pixel that the record time in the total pixel of background is earliest, it is ensured that the total pixel number amount of three class backgrounds is n；
If 3) be more than or equal to 10 with the current pixel quantity of the immediate background of new pixel, and new pixel color value with it is closest
The Euclidean distance of the average of background is less than 3 times of background standard difference, then new pixel is added in this classification background, and update this
The central point and standard deviation of classification background, delete record time earliest pixel in the total pixel of background, it is ensured that total picture of background
Prime number amount is n；
If 4) be not belonging to two kinds of situations above with the immediate background of new pixel, then it is assumed that be found that the background of new category；If
The background quantity at current time is equal to 3, then deletes that minimum class background of pixel quantity, and delete the institute for belonging to the background
There is pixel；After deletion, new classification background is set up by new pixel, new background central value is the color value of new pixel, new background
Pixel quantity be 1, standard deviation wouldn't be estimated.
7. a kind of automatic edit methods of multi-channel video suitable for campus football match according to claim 6, its feature
It is, the acquisition process of described moving target UNICOM body is that the new pixel for each frame occur is used as moving target candidate's picture
Element；For the moving target candidate pixel in each frame, the combination of pixels that all eight neighborhoods are connected is constituted multiple with this
Moving target UNICOM body.
8. a kind of automatic edit methods of multi-channel video suitable for campus football match according to claim 3, its feature
Be, the operation of described all moving target connected components of detection is, according to the area of described moving target UNICOM body and with
Round degree of approximation judges the current location of football, and according to the areal calculation current video of described moving target UNICOM body
The size of middle moving target.
9. a kind of automatic edit methods of multi-channel video suitable for campus football match according to claim 2, its feature
It is, the frame per second adjusting range of described video camera is 30-60.
10. a kind of automatic edit methods of multi-channel video suitable for campus football match according to claim 2, its feature
It is, the antenna height scope of described video camera is 2~4m.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710623659.6A CN107333031B (en) | 2017-07-27 | 2017-07-27 | Multi-channel video automatic editing method suitable for campus football match |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710623659.6A CN107333031B (en) | 2017-07-27 | 2017-07-27 | Multi-channel video automatic editing method suitable for campus football match |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107333031A true CN107333031A (en) | 2017-11-07 |
| CN107333031B CN107333031B (en) | 2020-09-01 |
Family
ID=60227694
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710623659.6A Expired - Fee Related CN107333031B (en) | 2017-07-27 | 2017-07-27 | Multi-channel video automatic editing method suitable for campus football match |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107333031B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108079556A (en) * | 2017-12-25 | 2018-05-29 | 南京云游智能科技有限公司 | A kind of universal self study coaching system and method based on video analysis |
| CN110049345A (en) * | 2019-03-11 | 2019-07-23 | 北京河马能量体育科技有限公司 | A kind of multiple video strems director method and instructor in broadcasting's processing system |
| CN111193961A (en) * | 2018-11-15 | 2020-05-22 | 索尼公司 | Video editing apparatus and method |
| CN111726649A (en) * | 2020-06-28 | 2020-09-29 | 百度在线网络技术（北京）有限公司 | Video stream processing method, device, computer equipment and medium |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101753852A (en) * | 2008-12-15 | 2010-06-23 | 姚劲草 | Sports event dynamic mini- map based on target detection and tracking |
| CN101777186A (en) * | 2010-01-13 | 2010-07-14 | 西安理工大学 | Multimodality automatic updating and replacing background modeling method |
| CN101795363A (en) * | 2009-02-04 | 2010-08-04 | 索尼公司 | Video process apparatus, method for processing video frequency and program |
| CN103959802A (en) * | 2012-08-10 | 2014-07-30 | 松下电器产业株式会社 | Video provision method, transmission device, and reception device |
| WO2015033546A1 (en) * | 2013-09-09 | 2015-03-12 | Sony Corporation | Image information processing method, apparatus and program utilizing a camera position sequence |
| CN105765959A (en) * | 2013-08-29 | 2016-07-13 | 米迪亚普罗杜申有限公司 | A Method and System for Producing a Video Production |
| CN106651952A (en) * | 2016-10-27 | 2017-05-10 | 深圳锐取信息技术股份有限公司 | Football detecting and tracking based video processing method and device |
-
2017
- 2017-07-27 CN CN201710623659.6A patent/CN107333031B/en not_active Expired - Fee Related
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101753852A (en) * | 2008-12-15 | 2010-06-23 | 姚劲草 | Sports event dynamic mini- map based on target detection and tracking |
| CN101795363A (en) * | 2009-02-04 | 2010-08-04 | 索尼公司 | Video process apparatus, method for processing video frequency and program |
| CN101777186A (en) * | 2010-01-13 | 2010-07-14 | 西安理工大学 | Multimodality automatic updating and replacing background modeling method |
| CN103959802A (en) * | 2012-08-10 | 2014-07-30 | 松下电器产业株式会社 | Video provision method, transmission device, and reception device |
| CN105765959A (en) * | 2013-08-29 | 2016-07-13 | 米迪亚普罗杜申有限公司 | A Method and System for Producing a Video Production |
| WO2015033546A1 (en) * | 2013-09-09 | 2015-03-12 | Sony Corporation | Image information processing method, apparatus and program utilizing a camera position sequence |
| CN106651952A (en) * | 2016-10-27 | 2017-05-10 | 深圳锐取信息技术股份有限公司 | Football detecting and tracking based video processing method and device |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108079556A (en) * | 2017-12-25 | 2018-05-29 | 南京云游智能科技有限公司 | A kind of universal self study coaching system and method based on video analysis |
| CN111193961A (en) * | 2018-11-15 | 2020-05-22 | 索尼公司 | Video editing apparatus and method |
| CN111193961B (en) * | 2018-11-15 | 2022-02-18 | 索尼公司 | Video editing apparatus and method |
| CN110049345A (en) * | 2019-03-11 | 2019-07-23 | 北京河马能量体育科技有限公司 | A kind of multiple video strems director method and instructor in broadcasting's processing system |
| CN111726649A (en) * | 2020-06-28 | 2020-09-29 | 百度在线网络技术（北京）有限公司 | Video stream processing method, device, computer equipment and medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107333031B (en) | 2020-09-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107333031A (en) | A kind of automatic edit methods of multi-channel video suitable for campus football match | |
| Scheerlinck et al. | CED: Color event camera dataset | |
| Stensland et al. | Bagadus: An integrated real-time system for soccer analytics | |
| Halvorsen et al. | Bagadus: an integrated system for arena sports analytics: a soccer case study | |
| US8922718B2 (en) | Key generation through spatial detection of dynamic objects | |
| US8451265B2 (en) | Virtual viewpoint animation | |
| EP2033140B1 (en) | Classifying image regions based on picture location | |
| US8154633B2 (en) | Line removal and object detection in an image | |
| CN105654471A (en) | Augmented reality AR system applied to internet video live broadcast and method thereof | |
| US8294824B2 (en) | Method and system for video compositing using color information in comparison processing | |
| CN110334635A (en) | Main body method for tracing, device, electronic equipment and computer readable storage medium | |
| KR101558467B1 (en) | System for revising coordinate in the numerical map according to gps receiver | |
| US20090129630A1 (en) | 3d textured objects for virtual viewpoint animations | |
| JP2018503301A (en) | System and method for displaying wind characteristics and effects in broadcast | |
| CN102739953B (en) | Image processing equipment, image processing method | |
| JP2020095717A (en) | Method, system and apparatus for capture of image data for free viewpoint video | |
| CN110049345A (en) | A kind of multiple video strems director method and instructor in broadcasting's processing system | |
| KR20160048178A (en) | A Method and System for Producing a Video Production | |
| CN101917546A (en) | Image processing apparatus and image processing method | |
| CN109712177A (en) | Image processing method, device, electronic equipment and computer readable storage medium | |
| CN103544696B (en) | A kind of suture line real-time searching method realized for FPGA | |
| CN108009491A (en) | A kind of object recognition methods solved in fast background movement and system | |
| CN110351579A (en) | A kind of intelligent editing algorithm of video | |
| Gruber et al. | UltraCamX, the large format digital aerial camera system by Vexcel Imaging/Microsoft | |
| CN101605269A (en) | A kind of method and apparatus of tracking dense depth images |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20200901 Termination date: 20210727 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |